Sensor sizes and f stops

Started Jan 25, 2014 | Discussions
Shop cameras & lenses ▾
starlight110
starlight110 Forum Member • Posts: 86
Sensor sizes and f stops

Hi, I know that in digital cameras, the bigger the sensor size the less noise and thus better light sensitivity, but I wonder if there´s a way to directly correlate sensor size and fstop, like doubling the sensor size from 1/1,8 to 1/3,6 would correlate to one fstop. I have an old low end A630 camera which has a 1/1,8 sensor size and performs average in low light, I want to buy a new more expensive camera but nowadays they make them all with 1/2,3 sensors and it would be weird to spend more money to buy a camera which performs (only from the sensor size factor) worse in low light. Would 1/1,8 to 1/2,3 size woiuld be much difference?

Thanks for the answers

trekkeruss
trekkeruss Veteran Member • Posts: 3,899
Re: Sensor sizes and f stops

There are compact cameras that have sensors bigger than 1/2.3"; lots of them, in fact. The most conspicuous are the Sony RX100 and RX100 mkII, which have 1" sensors. Then there is a whole crop of 1/1.7" cameras from Canon, Fujifilm, Nikon, Panasonic, Olympus. If you don't require a zoom lens, there are even compacts that have DSLR-size sensors. How large a sensor you can get depends only on how much money you are willing to pay.

Keep in mind that a wide aperture lens is still desirable, regardless of sensor size.

Reply   Reply with quote   Complain
OpticsEngineer
OpticsEngineer Senior Member • Posts: 4,397
Re: Sensor sizes and f stops

There are many camera similar in zoom range and basically the same chip size as your A630 but with much improved high ISO / low light performance due to better sensors and also faster lenses.

One place to look is the link below.

http://www.dpreview.com/articles/2367736880/roundup-enthusiast-zoom-compact-cameras

 OpticsEngineer's gear list:OpticsEngineer's gear list
Fujifilm XF1 Olympus XZ-2 iHS Canon PowerShot G7 X Olympus Tough TG-4 Sony Alpha NEX-7 +21 more
Reply   Reply with quote   Complain
darklamp
darklamp Senior Member • Posts: 3,567
f-numbers are independent of sensor size

I wonder if there´s a way to directly correlate sensor size and fstop

The way exposure system values work is that they are designed to give the same exposure on different systems, regardless of sensor size ( or film size ).

So there is no correlation between sensor size and f-number.

What is sometimes done is to calculate an equivalent f-number for a system to relate it to another. This is normally based on depth of field formula. However this ONLY relates to depth of field. You can also calculate the absolute size of the aperture from the f-number and the sensor size and compare them that way.

While there are small differences from system to system, using the same ISO, shutter speed and aperture (f-number) will result in the same exposure.

Reply   Reply with quote   Complain
Christof21
Christof21 Senior Member • Posts: 1,993
Re: f-numbers are independent of sensor size

darklamp wrote:

I wonder if there´s a way to directly correlate sensor size and fstop

The way exposure system values work is that they are designed to give the same exposure on different systems, regardless of sensor size ( or film size ).

So there is no correlation between sensor size and f-number.

What is sometimes done is to calculate an equivalent f-number for a system to relate it to another. This is normally based on depth of field formula. However this ONLY relates to depth of field. You can also calculate the absolute size of the aperture from the f-number and the sensor size and compare them that way.

While there are small differences from system to system, using the same ISO, shutter speed and aperture (f-number) will result in the same exposure.

Well, not really...

There is a correlation between sensor size and f-number. You can calculate an equivalent f number by multiplying by the crop factor. And it works for dof (same resulting dof) AND for the "light" ( same quantity of noise in the resulting image). You just have to stop down with the larger sensor and to raise the ISO to compensate.

The following link may help:

http://www.josephjamesphotography.com/equivalence/

Inthink this is the most misunderstood concept in photography....

So F1.8 with a P&S is like having F10 with a FF . See the difference ???

Reply   Reply with quote   Complain
darklamp
darklamp Senior Member • Posts: 3,567
No correlation

There is a correlation between sensor size and f-number.

No there isn't.

f-numbers are defined precisely to remove any correlation between aperture size ( i.e. the absolute measurement ) and the sensor.

You can calculate an equivalent f number by multiplying by the crop factor.

That's not a correlation.

http://en.wikipedia.org/wiki/Correlation_and_dependence

That's you defining another value using two others. Your new value has a dependence on sensor size, but the f-number you start from does not.

As I've already explained, that kind of equivalence is limited to depth of field.

Reply   Reply with quote   Complain
Christof21
Christof21 Senior Member • Posts: 1,993
Re: No correlation
1

darklamp wrote:

There is a correlation between sensor size and f-number.

No there isn't.

f-numbers are defined precisely to remove any correlation between aperture size ( i.e. the absolute measurement ) and the sensor.

You can calculate an equivalent f number by multiplying by the crop factor.

That's not a correlation.

http://en.wikipedia.org/wiki/Correlation_and_dependence

That's you defining another value using two others. Your new value has a dependence on sensor size, but the f-number you start from does not.

As I've already explained, that kind of equivalence is limited to depth of field.

I answer to the op who wonders what really means a f number when comparing with different sensor size.

Different sensor size should be taken into account, of course ! And you can calculate an "equivalent" aperture taking into account the crop factor.

There is a relation which "correlates" these equivalent apertures, I really think this is what the op wants to know.

It is very important to be aware of this. The op could buy a f1.8 with a1/2.3" thinking that it is better that a f2.8 with a 1" sensor. The 1" sensor is better for dof AND low light conditions.

Reply   Reply with quote   Complain
darklamp
darklamp Senior Member • Posts: 3,567
Re: No correlation
1

I'm not sure if English is your first language, but "correlate" does not mean what you're saying it does.

If there was a correlation between f-number and aperture you could find the one from the other, assuming other factors were constant.

You cannot.

You are just defining a value and as you factor in sensor size that defined value has a correlation with sensor size.

But the original f-number has NONE.

And if you start using the so-called equivalent f-number you have defined, you cannot use it for exposure calculations. So they are not equivalent in real exposures as you would have to use different shutter speeds and/or ISO values for different sensor sizes to get the same exposure.

Reply   Reply with quote   Complain
starlight110
starlight110 OP Forum Member • Posts: 86
Re: No correlation

Christof21 wrote:

It is very important to be aware of this. The op could buy a f1.8 with a1/2.3" thinking that it is better that a f2.8 with a 1" sensor. The 1" sensor is better for dof AND low light conditions.

That´s about what I ment with my original question, sorry if I didn´t express it correctly since my english is not the best and I´m newbie in photography.

I will put a simple example. Let´s say I want to compare two ideal cameras. Both have ideally fully manual controls and have the same lenses. One has a 1/2,3´sensor and the other 1/1,15´,, double in size, same megapixels and of the same type. I take a picture with the two cameras in the same light situation and setting in both iso100 and 1/30 sec exposure. The cameras have exactly the same algorithms to determine the aperture so that the amount of noise is about the same, in this case would there be a difference in aperture?? if yes, how much in terms of fstops??

Reply   Reply with quote   Complain
trekkeruss
trekkeruss Veteran Member • Posts: 3,899
Re: No correlation

If both cameras have the same shutter speed and ISO settings, the aperture should be the same as well; it does not matter what size is the sensor.

FWIW, if you are primarily interested in video, a video camera is often still a better choice. Video cameras usually have much wider aperture lenses than still cameras. They are also more ergonomic for shooting video.

Reply   Reply with quote   Complain
Christof21
Christof21 Senior Member • Posts: 1,993
Re: No correlation
2

starlight110 wrote:

Christof21 wrote:

It is very important to be aware of this. The op could buy a f1.8 with a1/2.3" thinking that it is better that a f2.8 with a 1" sensor. The 1" sensor is better for dof AND low light conditions.

That´s about what I ment with my original question, sorry if I didn´t express it correctly since my english is not the best and I´m newbie in photography.

I will put a simple example. Let´s say I want to compare two ideal cameras. Both have ideally fully manual controls and have the same lenses. One has a 1/2,3´sensor and the other 1/1,15´,, double in size, same megapixels and of the same type. I take a picture with the two cameras in the same light situation and setting in both iso100 and 1/30 sec exposure. The cameras have exactly the same algorithms to determine the aperture so that the amount of noise is about the same, in this case would there be a difference in aperture?? if yes, how much in terms of fstops??

No difference in aperture because the exposure rules are the same (triangle of exposure) whatever the sensor size is.

BUT you will have less noise with the larger sensor ! This is what is important. The difference is in fact 2 stops in favor of the larger sensor.

You could say that the larger sensor is the winner. Not that simple. In your example, you omit to say that when you take the same picture (same framing=same angle of view) , you will have less dof... So having this advantage of 2 stops is not free, you will have less dof.. If you stop down the larger sensor, you will have a similar picture in fact. With the  larger sensor, you have more choice in fact: you can get a similar picture or decide to have a shallower dof and more light. You do not have this choice with small sensors, it would need lenses with very low f number which do not exist.

Reply   Reply with quote   Complain
darklamp
darklamp Senior Member • Posts: 3,567
No real change in noise (SNR) in final image

The cameras have exactly the same algorithms to determine the aperture so that the amount of noise is about the same,

No. This will not work.

The camera will, using the same shutter speed and ISO and using the same metering system, will use the same f-number.

If you do not do this then one shot will be darker than the other and to equalize the shots you will have to brighten it. This is essentially equivalent to raising ISO and thus increasing noise relative to signal.

So you gain nothing by using a different aperture to equalize noise ( technically SNR ) instead of using the correct aperture.

Assuming you make both sensors the same way ( !! ) and they have the same number of pixel but different sizes, then the larger sensor should have less noise.

Because of the way noise is generated in small sensors I would expect the difference to be less than e.g. doubling the size might lead you to hope. However it's very hard to make generalizations in these cases.

I would suggest that the difference between older sensors and newer ones ( small sensors ) is also less dramatic than people like to think. Large sensors have improved a lot, but there are reasons for that which do not translate so well into small sensors.

Reply   Reply with quote   Complain
Christof21
Christof21 Senior Member • Posts: 1,993
Re: No real change in noise (SNR) in final image

darklamp wrote:

The cameras have exactly the same algorithms to determine the aperture so that the amount of noise is about the same,

No. This will not work.

The camera will, using the same shutter speed and ISO and using the same metering system, will use the same f-number.

If you do not do this then one shot will be darker than the other and to equalize the shots you will have to brighten it. This is essentially equivalent to raising ISO and thus increasing noise relative to signal.

So you gain nothing by using a different aperture to equalize noise ( technically SNR ) instead of using the correct aperture.

Assuming you make both sensors the same way ( !! ) and they have the same number of pixel but different sizes, then the larger sensor should have less noise.

It will not depend on the number of pixels.  You do not need to make this assumption. Noise does not depend on the pixel density.

Because of the way noise is generated in small sensors I would expect the difference to be less than e.g. doubling the size might lead you to hope. However it's very hard to make generalizations in these cases.

I would suggest that the difference between older sensors and newer ones ( small sensors ) is also less dramatic than people like to think. Large sensors have improved a lot, but there are reasons for that which do not translate so well into small sensors.

We can answer precisely to this by looking at the sensor efficiency. Dxomarks ratings is an excellent reference. Smaller sensors tend to be slightly more efficient but, in general, sensors efficiency does not vary with sensor size.

Reply   Reply with quote   Complain
darklamp
darklamp Senior Member • Posts: 3,567
Re: No real change in noise (SNR) in final image

Noise does not depend on the pixel density.

We'll be differing on this point, as I consider that a ridiculous statement which has been proven wrong repeatedly over the years in small sensor systems.

Reply   Reply with quote   Complain
TTMartin
TTMartin Veteran Member • Posts: 7,304
Re: No real change in noise (SNR) in final image

Christof21 wrote:

darklamp wrote:

The cameras have exactly the same algorithms to determine the aperture so that the amount of noise is about the same,

No. This will not work.

The camera will, using the same shutter speed and ISO and using the same metering system, will use the same f-number.

If you do not do this then one shot will be darker than the other and to equalize the shots you will have to brighten it. This is essentially equivalent to raising ISO and thus increasing noise relative to signal.

So you gain nothing by using a different aperture to equalize noise ( technically SNR ) instead of using the correct aperture.

Assuming you make both sensors the same way ( !! ) and they have the same number of pixel but different sizes, then the larger sensor should have less noise.

It will not depend on the number of pixels. You do not need to make this assumption. Noise does not depend on the pixel density.

Because of the way noise is generated in small sensors I would expect the difference to be less than e.g. doubling the size might lead you to hope. However it's very hard to make generalizations in these cases.

I would suggest that the difference between older sensors and newer ones ( small sensors ) is also less dramatic than people like to think. Large sensors have improved a lot, but there are reasons for that which do not translate so well into small sensors.

We can answer precisely to this by looking at the sensor efficiency. Dxomarks ratings is an excellent reference. Smaller sensors tend to be slightly more efficient but, in general, sensors efficiency does not vary with sensor size.

But, a larger sensor has more light reaching it. This can be illustrated with a full frame camera and and APS-C camera and the same full frame lens. A full frame Canon 6D camera has 864 square millimeters area that the light from the lens hits. An APS-C camera like the 700D has 332 square millimeters area that the light from the same lens hits. So even if both sensors have the same efficiency, the full frame camera will have over twice the performance, because it is collecting over twice the light.

 TTMartin's gear list:TTMartin's gear list
Canon PowerShot S100 (2000) Canon PowerShot SD1000 Canon PowerShot G1 X Canon EOS 40D Canon EOS 400D +26 more
Reply   Reply with quote   Complain
Christof21
Christof21 Senior Member • Posts: 1,993
Re: No real change in noise (SNR) in final image

TTMartin wrote:

Christof21 wrote:

darklamp wrote:

The cameras have exactly the same algorithms to determine the aperture so that the amount of noise is about the same,

No. This will not work.

The camera will, using the same shutter speed and ISO and using the same metering system, will use the same f-number.

If you do not do this then one shot will be darker than the other and to equalize the shots you will have to brighten it. This is essentially equivalent to raising ISO and thus increasing noise relative to signal.

So you gain nothing by using a different aperture to equalize noise ( technically SNR ) instead of using the correct aperture.

Assuming you make both sensors the same way ( !! ) and they have the same number of pixel but different sizes, then the larger sensor should have less noise.

It will not depend on the number of pixels. You do not need to make this assumption. Noise does not depend on the pixel density.

Because of the way noise is generated in small sensors I would expect the difference to be less than e.g. doubling the size might lead you to hope. However it's very hard to make generalizations in these cases.

I would suggest that the difference between older sensors and newer ones ( small sensors ) is also less dramatic than people like to think. Large sensors have improved a lot, but there are reasons for that which do not translate so well into small sensors.

We can answer precisely to this by looking at the sensor efficiency. Dxomarks ratings is an excellent reference. Smaller sensors tend to be slightly more efficient but, in general, sensors efficiency does not vary with sensor size.

But, a larger sensor has more light reaching it. This can be illustrated with a full frame camera and and APS-C camera and the same full frame lens. A full frame Canon 6D camera has 864 square millimeters area that the light from the lens hits. An APS-C camera like the 700D has 332 square millimeters area that the light from the same lens hits. So even if both sensors have the same efficiency, the full frame camera will have over twice the performance, because it is collecting over twice the light.

Interesting . Take the same full frame lens but use a speed booster adapter with the aps-c  (more than 1 stop gain) and the advantage vanishes !!!

Just the proof of equivalence.

Reply   Reply with quote   Complain
darklamp
darklamp Senior Member • Posts: 3,567
Light reaching each pixel

But, a larger sensor has more light reaching it. ...

You need to consider the number of photons reaching each pixel during an exposure ( not total area, but individual pixels ).

If I pack a large sensor with so many pixels that it has the same pixel density as the smaller sensor, then the resulting signal to noise ratio ( SNR ) will be the same.

Reply   Reply with quote   Complain
olliess
olliess Senior Member • Posts: 1,069
Re: Light reaching each pixel
1

darklamp wrote:

But, a larger sensor has more light reaching it. ...

You need to consider the number of photons reaching each pixel during an exposure ( not total area, but individual pixels ).

If I pack a large sensor with so many pixels that it has the same pixel density as the smaller sensor, then the resulting signal to noise ratio ( SNR ) will be the same.

If the larger sensor ends up with the same pixel density as the smaller sensor, then the larger sensor will have more total pixels.

Expose both sensors so that they receive the same number of photons per area... then the individual pixels from the two sensors will receive the same number of photons and they will have the same SNR, per pixel.

However, more pixels, and hence more photons, will be used to form any given area of the final image. The upshot being, the SNR normalized for the print area/resolution should still be better for the larger sensor.

Anyway this is getting deep into a rehash of stuff that's been rehashed endlessly. I wonder if the OP is still benefitting from this thread.  

Reply   Reply with quote   Complain
darklamp
darklamp Senior Member • Posts: 3,567
Re: Light reaching each pixel
1

However, more pixels, and hence more photons, will be used to form any given area of the final image. The upshot being, the SNR normalized for the print area/resolution should still be better for the larger sensor.

This is a common presumption people make, however noise does not average as nicely as people think it does.

As SNR gets lower averaging becomes much harder, to the point it's actually useless.

We can see this when we consider how effective reduced image sizes are at handling high ISO noise in some camera. Mush in, mush out, to put it crudely.

There is another side effect. SNR also limits DR. Small pixels lead to lower DR. You cannot average that out. So even if I accepted the idea that averaging worked well enough, the reduced DR is a killer.

Reply   Reply with quote   Complain
NetMage
NetMage Senior Member • Posts: 2,904
Re: Light reaching each pixel

More pixels is always better.

Nothing says you have to average the extra pixels - use advanced noise reduction that benefits from the extra pixels then reduce resolution - you get to pick the trade-off of detail and noise where fewer pixels force the loss of resolution, and doesn't reduce noise, just changes the scale.

-- hide signature --

-- Please remove the Quote option!
-- How about switching to real forum software?

 NetMage's gear list:NetMage's gear list
Canon PowerShot SD1000 Sigma DP1x Panasonic Lumix DMC-LX5 Fujifilm X10 Fujifilm X-S1 +23 more
Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads